Winners of the AI Safety Nudge Competition — LessWrong